Spatial and Survey Technology

A survey proposal is a written, formal document prepared before a survey is actually carried out. It states the problem to be studied, explains why the survey is needed, and lays out in advance what data will be collected, from whom, how, when, and at what cost. It is essentially the plan or blueprint of the survey, submitted to a funding agency, a supervisor, a research committee or a client for approval. Once approved, it becomes the working guide that keeps the survey on track and allows others to judge whether the study is feasible, relevant and methodologically sound.

Components of a survey proposal

1. Title of the survey — a short, clear and specific title reflecting the subject and scope.
2. Introduction / background — brief description of the topic, its context and the need for the study.
3. Statement of the problem — the exact question or gap the survey intends to address.
4. Objectives of the survey — general and specific objectives, stated in measurable terms.
5. Rationale / justification / significance — why the study is important, who will benefit, and its contribution to knowledge or policy.
6. Scope and limitations — geographical area, population, time period covered, and the constraints accepted.
7. Review of related literature — earlier studies and findings, showing the gap the present survey fills.
8. Research methodology — the core technical part, covering:
   • research design (descriptive, cross-sectional, etc.);
   • population and sampling design (sampling frame, size, technique);
   • data collection methods and instruments (questionnaire, interview schedule, observation);
   • variables and their measurement;
   • data processing and analysis techniques (editing, coding, tabulation, statistical tools).
9. Hypotheses / research questions (where applicable).
10. Work plan / time schedule — phases of the survey with expected dates (Gantt chart).
11. Budget / cost estimate — item-wise estimated expenditure and source of funding.
12. Expected outcomes / output — the form of the final report and its probable use.
13. References / bibliography — sources cited in the proposal.
14. Appendices — draft questionnaire, maps, tables, CV of the researcher, etc.

A good survey proposal is therefore complete, specific, realistic and clearly organised, so that the reader can understand the whole survey without its having been conducted.

Structured and Semi-Structured Instruments

Structured instrument: A structured (or standardized) instrument is a data-collection tool in which both the questions and the possible responses are fully predetermined before data collection. Every respondent receives exactly the same set of questions, in the same order and wording, and must choose from fixed response options. It is also called a closed-form or standardized instrument. Common examples are closed-ended questionnaires, rating scales (Likert scale), checklists, and structured interview schedules.

Semi-structured instrument: A semi-structured instrument is a data-collection tool in which the main questions are predetermined, but the researcher is free to probe further, rephrase, or add follow-up questions during the interview. It lies between the fully structured and the fully unstructured (open) instrument. A semi-structured interview guide or a questionnaire containing both closed-ended and open-ended questions are typical examples.

Characteristics of Structured Instruments

1. Predetermined questions and responses: Both the items and the answer choices are fixed in advance.
2. Standardization: The same instrument is administered to all respondents in an identical manner, which ensures uniformity.
3. Closed-ended items: Responses are mostly closed-ended (yes/no, multiple choice, rating scales), so answers are easy to code and quantify.
4. Quantitative in nature: The data collected are numerical or can be easily converted into numbers, making statistical analysis possible.
5. High reliability: Because of standardization, results are consistent and replicable across studies.
6. Low flexibility: The researcher cannot change questions or probe deeply; there is little room for follow-up.
7. Large sample and low cost: It can be administered to many respondents at once (especially by mail or online) and is relatively cheap and quick.
8. Impersonal: There is little personal contact between researcher and respondent, so rapport is limited.
9. Suitable for descriptive and survey research: Best when the researcher already knows what to measure and wants to test hypotheses on a large scale.

Characteristics of Semi-Structured Instruments

1. Partly predetermined: A core set of questions or themes is prepared in advance, but the order and wording may vary.
2. Flexibility: The researcher can probe, clarify, and ask follow-up questions to obtain deeper information.
3. Mix of closed and open-ended items: Some questions have fixed options while others allow free responses.
4. Qualitative plus quantitative data: Produces both numerical data and rich, descriptive information.
5. Moderate reliability and high validity: Less standardized than structured tools (lower reliability), but flexibility allows deeper and more valid exploration of complex issues.
6. Requires skilled interviewer: The quality of data depends heavily on the interviewer's skill in probing and guiding the conversation.
7. Smaller sample, higher cost and time: Usually administered to fewer respondents through interviews, so it is more time-consuming and expensive.
8. Rapport and personal contact: Direct interaction helps build trust and encourages honest, detailed responses.
9. Suitable for exploratory and explanatory research: Useful when the researcher has some prior knowledge but also wants to discover new insights or understand reasons behind responses.

Conclusion

Structured instruments are rigid, standardized, and best for large-scale quantitative surveys where reliability and comparability matter. Semi-structured instruments are partly flexible and best for studies that need both breadth and depth. The choice between them depends on the research objectives, nature of the problem, sample size, and available resources.

Importance of Web-Based Questionnaires

A web-based questionnaire is a survey instrument that is designed, distributed, and filled in over the internet — the respondent opens a link in a browser and submits answers directly to a server. It has become the dominant mode of data collection in research, business, and government because it removes the need for paper, postage, printing, and physical presence of an enumerator. It is important because it allows a researcher to reach a large, geographically scattered population quickly and at very low cost, it collects responses in an already digital form so that coding, tabulation, and statistical analysis can begin immediately, and it can enforce consistency (skip logic, mandatory fields, range checks) so that errors and missing values are reduced at the point of entry. It also makes large-scale, repeated, and longitudinal surveys practical, and during situations such as the COVID-19 pandemic it was often the only feasible mode of data collection.

Advantages

1. Low cost — no printing, postage, stationery, or travel; a single form can serve thousands of respondents.
2. Wide reach and speed — respondents anywhere in the world can be reached within minutes through email or social media, and results are available in real time.
3. Automatic data entry — responses are stored directly in a database or spreadsheet, eliminating manual data entry and its transcription errors.
4. Design flexibility — multimedia (images, audio, video), branching/skip logic, randomisation of options, and personalised questions are possible.
5. Convenience for respondents — they can answer at a time and place of their choice, and can pause and resume.
6. Anonymity and honesty — for sensitive topics (income, health, sexual behaviour, corruption) respondents tend to answer more truthfully than in face-to-face interviews.
7. Easy follow-up and monitoring — reminders can be sent automatically, and completion rates and drop-off points can be tracked.
8. Large samples and easy analysis — data can be exported directly to SPSS, Excel, R, or Python for statistical analysis.

Disadvantages

1. Limited population coverage — only those with internet access, a device, and digital literacy can participate; this excludes the poor, the elderly, and rural populations, causing coverage bias.
2. Low response rate — people often ignore or delete survey links; non-response bias may make the sample unrepresentative.
3. Self-selection bias — those who choose to respond may differ systematically from those who do not.
4. No interviewer control — the researcher cannot clarify ambiguous questions, probe incomplete answers, or verify who actually filled the form.
5. Technical problems — server downtime, broken links, browser incompatibility, and page-load failures can lose responses.
6. Security and privacy risks — hacking, data leaks, and unauthorised access to personal data; ethical consent is harder to ensure.
7. Superficial responses — respondents may rush, skip questions, or give careless answers because there is no one to motivate them.
8. Design constraints — complex or long questionnaires are difficult; poorly designed forms increase drop-out.

Conclusion: Web-based questionnaires are important because they make large-scale, low-cost, fast, and error-reduced data collection possible. Their main strength is cost and reach; their main weakness is coverage and response bias. They are therefore best used together with other methods (telephone or face-to-face interviews) when the target population is not fully online.

Quantitative Data Source

A quantitative data source is any source that provides data in numerical or measurable form — counts, measurements, rates, ratios or amounts — which can be tabulated, averaged and analysed statistically. Examples include population censuses, national accounts, sample surveys, price indices and administrative registers. Such data answer questions of how much, how many or how often, and are the basis for descriptive statistics, comparison over time or place, and mathematical modelling.

Can administrative data be considered a quantitative data source?

Yes. Administrative data is the information routinely collected and maintained by government agencies, institutions or organisations while carrying out their normal functions — for example, birth and death registers, tax records, school enrolment records, hospital admission records, land registration records, passport and citizenship records, and customs records. Since these records consist mainly of numbers (number of births, amount of tax collected, number of students enrolled, number of patients treated), they are a genuine quantitative data source. They are sometimes called "by-product" data because they are generated for administrative purposes, not primarily for research. Their quantitative character is beyond doubt; the only caution is that they are collected for control and record-keeping, so their definitions, coverage and quality may not match the needs of a particular study.

Purposes of using administrative data as a quantitative source

1. Low cost and quick availability — the data already exists, so no separate costly survey or field enumeration is needed.
2. Complete coverage of the target population — since registration or recording is compulsory, coverage is often near-universal (e.g. all births registered in a district).
3. Continuous and timely series — records are updated regularly, giving a continuous flow of data for monitoring trends.
4. Small-area and disaggregated analysis — data can be broken down by ward, district, sex, age, caste/ethnicity, etc., which is costly in sample surveys.
5. Benchmark and sampling frame — used to check the reliability of survey estimates and to provide the sampling frame for sample surveys.
6. Planning, monitoring and evaluation — supports policy formulation, budget allocation, programme monitoring and impact evaluation by the government and other agencies.
7. Reduced respondent burden — no repeated questioning of the same respondents, as the information is already collected during service delivery.

Conclusion

Administrative data is therefore a valid and valuable quantitative data source. Its main limitations — under-reporting, definitional mismatch, and dependence on the quality of the recording system — must be checked before use, but when handled carefully it is one of the cheapest and most useful sources for statistical analysis.

Resolution in Remote Sensing

Resolution refers to the ability of a remote sensing system to distinguish (separate) two closely spaced objects or to record fine details of a target. In other words, it is the smallest level of detail that a sensor can detect and represent in the resulting image or data. A higher resolution means finer detail can be observed, while a lower resolution means only coarse features are visible.

Types of Resolution

Remote sensing resolution is generally classified into four main types:

1. Spatial Resolution It is the ground area represented by a single pixel (picture element) of an image. It indicates the smallest object that can be distinguished on the ground.

• Example: A spatial resolution of 30 m means one pixel covers a 30 m × 30 m ground area (as in Landsat TM).
• High spatial resolution (e.g., 0.5 m in IKONOS/QuickBird) shows fine details; low spatial resolution (e.g., 1 km in AVHRR) shows only broad features.

2. Spectral Resolution It refers to the number and width of wavelength bands (spectral intervals) in which a sensor records radiation. A sensor with many narrow bands has high spectral resolution.

• Example: Multispectral sensors (Landsat: 7 bands) have moderate spectral resolution; hyperspectral sensors (e.g., AVIRIS: 200+ narrow bands) have very high spectral resolution.

3. Temporal Resolution It is the time interval (revisit period) between two successive images of the same area taken by a sensor.

• Example: Landsat has a 16-day revisit period; MODIS has daily coverage. High temporal resolution is useful for monitoring dynamic phenomena such as floods, crops and cyclones.

4. Radiometric Resolution It refers to the sensitivity of a sensor to detect small differences in the intensity (brightness) of radiation. It is expressed in terms of the number of bits or grey levels used to record data.

• Example: An 8-bit sensor records 2⁸ = 256 grey levels; a 12-bit sensor records 2¹² = 4096 grey levels. Higher radiometric resolution means finer discrimination of radiance differences.

Conclusion Thus, resolution in remote sensing is the measure of a sensor's capability to resolve detail. Spatial, spectral, temporal and radiometric resolutions together determine the suitability of a sensor for a particular application, and a trade-off often exists among them.

Similarities between Vector and Raster Data Models

1. Both represent spatial data — Each model stores geographic features (location, shape, and attributes) in a digital form that a GIS can display and analyse.
2. Both use a coordinate system — Data in either model is georeferenced to a common coordinate system (geographic or projected) so that layers can be overlaid.
3. Both store attributes — In both cases, geometric/spatial data is linked to non-spatial attribute data (usually through an attribute table or value codes).
4. Both support spatial analysis — Operations such as overlay, buffering, clipping, and querying can be performed on either type of data.
5. Both are used together in a GIS — A single GIS project commonly combines vector and raster layers (e.g., a vector road layer over a raster satellite image).
6. Both can be converted — Raster can be vectorised and vector can be rasterised, so the two models are interchangeable to some extent.

Differences between Vector and Raster Data Models

Conclusion: The vector model is object-based and suited to discrete, well-defined features, while the raster model is cell-based and suited to continuous phenomena. Both are complementary, and a complete GIS usually employs both.

Variable selection (also called feature selection or subset selection) is the process of choosing the most relevant independent variables (predictors) from a larger set of available variables for inclusion in a regression (or other statistical) model. The aim is to keep only those variables that contribute significantly to explaining the variation in the dependent variable, while discarding irrelevant, redundant, or multicollinear ones. This gives a simpler, more interpretable model with better predictive accuracy and more reliable coefficient estimates.

How the process is done

1. Define the problem and the dependent variable — decide what is to be explained or predicted.
2. List all candidate independent variables — collect every variable that may plausibly affect the dependent variable, based on theory or prior study.
3. Examine the data — check for missing values, outliers, and correlations among predictors (multicollinearity).
4. Apply a selection criterion — fit models and judge each variable by statistical measures such as the coefficient of determination (R²), adjusted R², t-test (p-value), F-test, Akaike Information Criterion (AIC), or Mallows' Cₚ.
5. Select the final subset — retain variables that meet the chosen criterion and are theoretically meaningful; drop the rest.
6. Validate the model — check residual diagnostics and, where possible, confirm performance on fresh data.

Ways (methods) of variable selection

1. Forward selection — start with no variables; add one variable at a time, each time the one that most improves the criterion, and stop when no further significant improvement occurs.
2. Backward elimination — start with all candidate variables; remove the least significant variable one at a time until all remaining variables are significant.
3. Stepwise selection — a combination of the two: at each step a variable may be added, and previously entered variables may be removed if they become insignificant.
4. All possible subsets (best subset) regression — fit every possible combination of predictors and choose the subset with the best value of the criterion (e.g., highest adjusted R² or lowest AIC/Cₚ).
5. Ridge regression and LASSO — shrinkage methods that penalise large coefficients; LASSO can shrink some coefficients exactly to zero, thereby performing selection automatically.
6. Principal component regression / factor analysis — reduce many correlated variables into a few uncorrelated components and use those as predictors.
7. Judgemental (a priori) selection — variables are chosen on the basis of theory, past experience, or expert opinion rather than a purely statistical search.

Major Problems in Survey Research

Survey research, though widely used, suffers from several major problems:

1. Problem of identifying the respondent — It is often difficult to locate and identify the actual person who can give reliable information (e.g., the real head of household, the actual decision-maker, or the genuine consumer).
2. Non-response / refusal — Many selected respondents refuse to answer or are absent, leading to non-response bias and a reduced effective sample.
3. Sampling and coverage error — The sampling frame may be incomplete or outdated, so some elements of the population have no chance of selection.
4. Response bias — Respondents may give socially desirable answers, exaggerate, under-report, or deliberately mislead.
5. Recall and memory error — Respondents may not remember past events accurately, especially for events long past.
6. Question and questionnaire defects — Ambiguous, leading, double-barrelled or poorly worded questions distort responses.
7. Interviewer bias — The interviewer's appearance, tone, or expectations may influence the respondent's answers.
8. Cost, time and administrative difficulty — Surveys are expensive, time-consuming, and require trained enumerators and supervision.
9. Difficulty in establishing causality — Surveys describe "what is" but cannot easily prove cause-and-effect relationships.
10. Lack of control — The researcher cannot control extraneous variables as in an experiment.

How the Problem of Identifying the Respondent Can Be Reduced

The problem of identifying the correct respondent can be reduced by:

1. Preparing a clear sampling frame — Listing all units of the population with addresses and identification details before the survey.
2. Defining the respondent precisely — Clearly specifying who qualifies as the respondent (e.g., "household head", "person above 18 years", "principal buyer").
3. Using screening questions — Asking filter questions at the beginning to confirm that the person meets the required criteria.
4. Household listing / listing schedule — First listing all members of a household and then selecting the eligible respondent by a random or Kish-grid method.
5. Proper training of enumerators — So they can verify identity, relationship and eligibility before interviewing.
6. Using identification aids — Maps, voter lists, citizenship records, telephone directories and GPS locations to trace respondents.
7. Repeated visits and call-backs — To contact absent or temporarily unavailable respondents.
8. Pilot testing — To check whether the identification procedure works in the field.
9. Community contact persons — Local leaders or informants who help locate the right respondent.
10. Maintaining respondent records — Keeping a register of contacted, identified and interviewed respondents to avoid duplication.

Advantages and Disadvantages of Survey Compared with Other Methods

Advantages over other methods

1. Large and representative sample — Surveys can cover a wide population, giving results that can be generalized (unlike case study or observation, which cover few cases).
2. Standardized measurement — The same questions are put to all respondents, making comparison and quantification easy.
3. Collection of both facts and opinions — Surveys can measure attitudes, beliefs, preferences and behaviour, which observation alone cannot capture.
4. Economy and speed — Sample surveys are cheaper and faster than census or long-term experiments.
5. Flexibility — Can be conducted by personal interview, telephone, mail or online, suiting different situations.
6. Statistical analysis possible — Data can be tabulated and analysed using statistical tools, and inferences drawn about the population.
7. Anonymity — In mail or online surveys, respondents may reveal sensitive information more freely than in face-to-face interviews.
8. Replicability — A survey can be repeated to check reliability and to study trends over time.

Disadvantages compared with other methods

1. Superficiality — Surveys give broad but shallow information; they cannot probe deeply into motives as interviews or case studies can.
2. Inflexibility of schedule — Once the questionnaire is fixed, questions cannot be changed during the survey (unlike unstructured interviews).
3. Dependence on respondent's ability — Success depends on the respondent's memory, honesty and willingness, which experiments and observation do not require.
4. Non-response and bias — High refusal rates reduce representativeness, a problem less serious in observation or documentary methods.
5. No control over variables — Unlike experiments, surveys cannot manipulate independent variables to establish causation.
6. Costly for large areas — A nationwide personal-interview survey is more expensive and time-consuming than a laboratory experiment or a small case study.
7. Risk of artificiality — Respondents may answer in an artificial setting, not reflecting real behaviour; observation records actual behaviour.
8. Difficulty in verification — Responses cannot be easily verified, whereas documentary or observational data can be cross-checked.

Conclusion: Survey is the most suitable method when the aim is to describe a large population's characteristics, attitudes or behaviour in a standardized, quantifiable way. Its weaknesses — respondent identification, non-response and superficiality — can be reduced through careful sampling-frame preparation, clear respondent definition, screening questions and trained enumerators. For deep causal study, however, experiment or case study remains superior.

Remote Sensing

Remote sensing is the science and art of acquiring information about an object, area, or phenomenon without making physical contact with it. The information is obtained by sensing and recording reflected or emitted electromagnetic (EM) energy, which is then processed and interpreted to extract useful data about the target. A typical remote sensing system consists of a source of energy, a sensor, a platform (aircraft or satellite), and a data-processing/interpretation chain.

Physical Basis of Remote Sensing

The physical basis of remote sensing rests on the interaction of electromagnetic radiation with matter. Its key elements are:

1. Electromagnetic radiation as the carrier of information — All matter above absolute zero emits EM radiation. Remote sensing uses this radiation as the medium that carries information from the target to the sensor. EM radiation travels in vacuum at the speed of light and behaves as both a wave and a particle (photon).

2. Wave parameters — Radiation is described by wavelength (λ), frequency (ν), and energy (Q). They are related as:

$c = \nu \lambda \qquad \text{and} \qquad Q = h\nu = \frac{hc}{\lambda}$

where c = speed of light (3 × 10⁸ m/s) and h = Planck's constant (6.626 × 10⁻³⁴ J·s). Thus shorter wavelengths carry higher energy.

3. Electromagnetic spectrum — Remote sensing operates in selected atmospheric windows within the spectrum: visible (0.4–0.7 µm), near- and mid-infrared (0.7–3 µm), thermal infrared (3–14 µm), and microwave (1 mm–1 m). Each region reveals different target properties.

4. Energy–matter interaction — When incident energy (Eᵢ) strikes a target, it is partitioned into reflected (Eᵣ), absorbed (Eₐ), and transmitted (Eₜ) components:

$E_i = E_r + E_a + E_t$

The reflected portion is what most optical sensors measure. Reflectance varies with wavelength, giving each object a spectral signature (spectral reflectance curve) that enables identification.

5. Radiation laws — Planck's law gives spectral radiant emittance of a blackbody; Stefan–Boltzmann law states total emitted energy M = σT⁴; Wien's displacement law gives λ_max = 2898/T (µm), explaining why hotter bodies emit at shorter wavelengths. These laws govern thermal remote sensing.

6. Atmospheric effects — Radiation must pass through the atmosphere, where it is scattered (Rayleigh, Mie, non-selective) and absorbed by gases (O₃, CO₂, water vapour). Only atmospheric windows allow useful sensing.

Operation of Energy Sources and Radiation Principles

A. Energy sources

• Passive remote sensing — The sensor relies on energy available naturally. The ultimate source is the Sun, which radiates energy peaking in the visible region (~0.48 µm) according to Planck's law. Sunlight illuminates the target; the sensor records reflected solar radiation (visible–reflected IR) or emitted terrestrial radiation (thermal IR). Examples: Landsat, SPOT, aerial photography.
• Active remote sensing — The system carries its own artificial energy source (e.g., radar antenna, LiDAR laser) that transmits a pulse toward the target and measures the returned signal. This allows day/night and all-weather operation. Examples: RADARSAT, LiDAR, SLAR.

B. Radiation principles in operation

1. Emission — The Sun (≈6000 K) emits short-wave radiation; the Earth (≈300 K) emits long-wave thermal IR. Wien's law fixes the spectral region of emission for each.
2. Propagation and atmospheric transmission — Energy travels to the target through the atmosphere; sensors are tuned to atmospheric windows to minimise absorption and scattering losses.
3. Interaction with the target — The target reflects, absorbs, or transmits energy depending on its physical and chemical properties. Reflectance in different bands produces the spectral signature used for discrimination (e.g., healthy vegetation reflects strongly in NIR, water absorbs NIR).
4. Re-radiation and return path — Reflected or emitted energy travels back through the atmosphere to the sensor.
5. Sensing and recording — The sensor converts incoming radiant flux into an electrical signal, which is digitised into brightness values (DNs) per pixel per band.
6. Processing and interpretation — Radiometric and geometric corrections, enhancement, classification, and visual interpretation convert the recorded radiance into thematic information (land cover, temperature, moisture, etc.).

In short, remote sensing works because every object interacts with EM radiation in a characteristic way; by controlling or observing the energy source and applying radiation laws, the sensor captures this interaction and converts it into interpretable information about the Earth's surface.

Physical and Functional Components of GIS

A Geographic Information System (GIS) is best understood through two complementary views: its physical components (the tangible parts that make up the system) and its functional components (the operations or capabilities it performs on geographic data).

1. Physical Components of GIS

These are the actual hardware, software, data, people and procedures that together form a working GIS.

(a) Hardware The computer system on which GIS runs. It includes:

• Computers — desktop workstations, laptops or servers that process spatial data.
• Input devices — digitizers, scanners, GPS receivers, total stations and keyboards used to capture data.
• Output devices — plotters, printers, monitors and screens for displaying maps and reports.
• Storage devices — hard disks, SSDs, servers and cloud storage for holding large spatial datasets.

(b) Software The programs that provide GIS tools and functions. It includes:

• GIS packages — ArcGIS, QGIS, GRASS GIS, ERDAS Imagine.
• Database management software — PostgreSQL/PostGIS, Oracle Spatial, SQL Server.
• Supporting software — operating systems, image processing tools and programming libraries (Python, GDAL).

(c) Data The most valuable and expensive component. GIS data consists of:

• Spatial data — location information in vector (point, line, polygon) or raster (grid cell) form.
• Attribute data — descriptive, non-spatial information linked to spatial features (e.g., population, land use).
• Metadata — data about data, describing source, accuracy, projection and date.

(d) People The users who design, operate and interpret the GIS — analysts, planners, surveyors, decision-makers and technicians. People are the most critical component because they turn data into useful information.

(e) Methods / Procedures The rules, standards and workflows that govern how data is collected, stored, analysed and presented. These ensure consistency, accuracy and reproducibility of results.

2. Functional Components of GIS

These describe what a GIS does — the sequence of operations performed on geographic data.

(a) Data Input and Capture Acquisition of spatial data from maps, field surveys, GPS, remote sensing images or existing databases. Includes digitizing, scanning, georeferencing and data conversion.

(b) Data Storage and Management Organizing spatial and attribute data in databases (geodatabases, shapefiles, raster catalogs) so they can be efficiently retrieved, updated and maintained.

(c) Data Manipulation and Analysis The core strength of GIS. Includes:

• Overlay analysis — combining layers (union, intersect, clip).
• Buffer analysis — creating zones around features.
• Network analysis — route finding, service areas.
• Spatial interpolation — estimating values at unknown points.
• Query and statistics — selecting features by location or attribute.

(d) Data Output and Visualization Presenting results as maps, graphs, charts, tables or 3D models for interpretation and decision-making.

(e) Data Updating and Maintenance Continuous correction, revision and updating of data to keep the GIS current and reliable.

---

Method of Defining Projection and Coordinate System in GIS

Every spatial dataset must be tied to a coordinate system and a map projection so that locations on the curved Earth can be represented on a flat map. In GIS software (e.g., ArcGIS, QGIS), this is done through the following steps.

Step 1: Understand the Two Types of Coordinate Systems

(a) Geographic Coordinate System (GCS)

• Uses a spheroid/datum (e.g., WGS 1984, Everest 1830) to model the Earth.
• Locations are expressed in latitude and longitude (angular units — degrees, minutes, seconds).
• It is a 3D spherical reference system.

(b) Projected Coordinate System (PCS)

• Built on a GCS plus a map projection (e.g., UTM, Lambert Conformal Conic).
• Locations are expressed in linear units (metres, feet) on a flat surface.
• It is a 2D planar reference system.

Step 2: Select the Appropriate Datum

The datum defines the size, shape and position of the reference ellipsoid relative to the Earth.

• Global datum: WGS 1984 (used by GPS).
• Regional datum: Everest 1830 or Nepal datum for Nepal. Choosing the correct datum is essential; a wrong datum causes positional errors of hundreds of metres.

Step 3: Choose a Suitable Map Projection

The projection is chosen based on:

• Purpose — whether area, shape, distance or direction must be preserved.
• Extent — local, regional or global.
• Location — latitude and longitude of the study area.

Common projections:

• Transverse Mercator / UTM — for local and regional mapping.
• Lambert Conformal Conic — for mid-latitude regions.
• Albers Equal Area — when area accuracy is important.

Step 4: Define the Coordinate System in GIS Software

In ArcGIS:

1. Open ArcCatalog or the Data Management Tools → Projections and Transformations toolbox.
2. Right-click the dataset → Properties → XY Coordinate System.
3. Choose Geographic or Projected coordinate system from the predefined list (e.g., GCS_WGS_1984, WGS_1984_UTM_Zone_45N).
4. Alternatively, use Define Projection tool to assign a coordinate system to data that lacks one.
5. Use Project tool to transform data from one coordinate system to another.

In QGIS:

1. Go to Project → Properties → CRS.
2. Search and select the required CRS (e.g., EPSG:4326 for WGS 84, EPSG:32645 for UTM Zone 45N).
3. For individual layers: right-click layer → Properties → Source → Assign CRS.
4. Use Vector → Data Management Tools → Reproject Layer to change projection.

Step 5: Verify and Validate

• Check the metadata for correct projection information.
• Overlay the layer with a known reference layer (e.g., Google Earth, base map) to confirm alignment.
• Ensure all layers in a project share the same coordinate system to avoid misalignment during overlay analysis.

Step 6: Document the Projection

Record the projection name, datum, units and EPSG code in the metadata so that the data remains usable and interpretable by other users.

---

Conclusion: The physical components (hardware, software, data, people, methods) provide the infrastructure of a GIS, while the functional components (input, storage, analysis, output, updating) describe its working capability. Correctly defining the projection and coordinate system is fundamental, because spatial analysis and mapping are meaningful only when every feature is referenced to a common, accurate geographic framework.